Logged RPM Wobbles

[Cus]

Greeting Humans!

Reading through another thread on an entirely different topic, I saw Magpie had posted the following graph:

TPS_01 by Eipeip, on Flickr



Notice the yellow line (RPM) on the right where he changes gear - there's an RPM "wobble" - now the reason I've noticed is because my car does the same thing:


White line is RPM on my graph, and my wobble is far more pronounced. The reason I've noticed it in the first place is it's noticeable in the car as a jerky motion when changing gears while "on it" (as far as one can be on it in a stock 1.6) but the initial up swin is 450RPM, and then it gets pulled back 200RPM, then it wings back up 250, down 110, then starts being within the noise margin.

Initially I thought it might have been tune related because the injector duty cycle follows the RPM pretty closely (and I'm pretty clueless about tuning), but lately I've been thinking diff bushes or engine mounts - or both - because I've noticed the car can be rocked back and forward slightly while the handbrake is on. (nothing is loose, I've checked that much at least!)

And now I see someone else's car does it, and I know for a fact Magpie's car has had far more intelligent people working on the tune than mine has

So, Question 1:
Has anyone else noticed this?

Question 2:
Have you solved it?

Question 3:
How "solved" can this expected to be, given the design of the MX5 diff mounting?

*Casts Summon Magpie*

Out of curiosity, how old are the bushes in your diff, and possibly engine mounts?


Why do I care? Well, I don't really care at this point, but I think It might go some way towards explaining why some NA6 diffs blow to bits really easy, while some seems to be slightly tougher; ie: the drivetrain is "winding up" and then relaxing several times as opposed to just winding up and putting power to the ground due to - i'm guessing - worn diff bushes.

[StanTheMan]

well I can only say.....NA diffs blow very easy at the best of times.

Mine was just accelerating gently in traffic at the time. On a stock motor with a few bolt ons.

It seems as though from reports the guys racing seem to blow diffs a lot less than the road ones. I do believe some LSD diffs make it slightly stronger. Whiu;le perhaps in traffic you tend to be more jerky more often while in racing maybe its always WOT or not. Perhaps your right.....think about upgrading your NA6 diff to a NA8 or NB one

[Magpie]

Could be a couple of issues.

1. The frequency of logged data. I'll find out what the RPM log rate is when I get home.
2. Noise. There is similar 'wobbles' on the other lines with the exception of the TPS (green)
3. The scale being used, this coupled with logging rate will cause 'bumpy' lines.
4. Tune related as fuel is injected because of acceleration. I don't feel the 'wobbles', unless it is a failing CAS sensor...

This is NOT what you want to see! The blue line is the RPM of a failing Cam Angle Sensor
Miss EC Day 1 S3 by Eipeip, on Flickr

My diff and engine mounts are about 2 years old.

[Roadrunner]

I've noticed this in my logs when accelerating over dips and bumps in the track.
Below is coming onto the main straight at Marulan, you can see the RPM slightly wobble when the clutch is engaged. The track isn't the smoothest here including track camber change and then halfway through the pull is where the new track crosses the old track and theres quite a big bounce in the RPM. This is 3rd gear and I can feel this bounce in the car and in RPM/engine load as I cross the dip where the two tracks meet.

I always assumed this wobble is just the car bouncing slightly when changing gear (IE: Off throttle clutch in - weight transfer causes nose to dip - get back on throttle - weight transfer causes rear suspension to squat then neutralise soon after but puts oscillating load on the rear till theres no more bounce) or rough track patches.

[Magpie]

I thought the same, bumps but the TPS is at 100% so it is not a 'foot' issue.

[Roadrunner]

Could it just be different load amounts on the rear wheels while at full throttle? Ie when the dip/bump is hit the rear squats and the RPM drops slightly as there's more weight on the drive wheels/diff etc, then when it bounces back up there's less weight on the rear allowing the engine to increase RPM faster, and then RPM bogs back down when the weight loads up on the rear again. All in rather small amounts, though with standard springs/shocks I'd imagine this weight/load transfer would be more pronounced, all while at 100% TPS.

Imagine on a bigger (hypothetical) scale, at 100% TPS you hit a bump in the road that lifts the rear enough to break traction. This would show as a huge oscillation in RPM as traction is lost (RPM increases dramatically) and as the car comes back down and the drive wheels get grip again, the RPM would drop dramatically and be lower than it should be due to more weight on the back/drive wheels. This oscillation in RPM will smooth out as the suspension smooths out the vehicle bounce and will then be linear as the suspension is now neutral (not bouncing). All happening while at full throttle.

The way to test this theory would be to see if anyone has a data logger than measures suspension compression etc to compare with the RPM bounce and see if it lines up.

[Magpie]

Dammm, I have some logs with rear suspension and RPM so will have a look.

[Cus]

1. The frequency of logged data. I'll find out what the RPM log rate is when I get home.
Logging speed on my ECU is 100Hz, one log event every 10ms, one sine wave in my logs is about 25 log events (250ms)

2. Noise. There is similar 'wobbles' on the other lines with the exception of the TPS (green)

3. The scale being used, this coupled with logging rate will cause 'bumpy' lines.
"Yes, but" - Some inputs seem very clean, others are very noisy, my VSS input is all over the place like a madman's sh*t.

4. Tune related as fuel is injected because of acceleration. I don't feel the 'wobbles', unless it is a failing CAS sensor...
Being two oscillations in that example, the first one would be lost in the feel of changing gear, and the second one probably isn't big enough to notice.

2+3:My Injector Duty Cycle and Spark Advance for example will follow the line, DC almost exactly, but if the ECU is reacting to the RPM of the engine, this would be the correct behaviour (ie; symptom, not a cause)

I've noticed this in my logs when accelerating over dips and bumps in the track.

Your 920 line goes almost exactly through a "normal" wobble, but you have one pronounced oscillation, and then some much smaller ones like a logarithmic drop-off, where as Magpie's car looks like more of a linear drop-off but much faster than mine. The bumps in the middle do look like the drivetrain loading up / unloading as the wheel gets heavier and lighter over the bumps, and look to be approximately the same amplitiude as a gear-change wobble. This makes me think "bushes".

My gear change looks to have been absolutely terrible - there is 350ms between throttle (green line) starting to open and the RPM getting to it's lowest position (clutch fully released) So I'm not going to rule out the issue being PEBKAC, but for cars (Problem Exists Between Keyboard and Chair)

Having said that, I dug out the logs from my drive down the Oxley, and I've actually got decel wobbles too. Of a full 16 minute log I went WOT once (which was passing the only bit of traffic for the whole run!) most of the throttle was below 50%

The car is physically doing 92ish Km/h here, throttle is 15%.


But both of you guys have fewer oscillations than I do, here's an extreme example: (3rd gear, going from 8% to 12% throttle, following someone down the Oxley prior to overtaking them)

Yes, that's 1.7 seconds of RPM oscillations with light throttle tip-in at 45km/h. I'm pretty sure that's not suspension at play, I've got 6kg springs in the rear and I'm not even close to being "into it" on this one, and the Oxley is pretty smooth.

It does seem to me like it's a diff bushes issue, in that mine are basically "failed". What are you guys using for bushes, and how old are they?

The other side of the coin is that Roadrunner having a turbo (probably) has more power than Magpie, and Magpie with a built 1.8 has more power than my stock 1.6. The theory there being the amount of power being transmitted through the drive train is far more than the drivetrain can "push back" once it's loaded up, therefore less oscillations.

I really hope that makes sense. It does in my head, but weird things happen in there....

I'd be curious to see what Magpie's suspension logs look like regardless.

[speed]

Maybe you need to consult with the wobble master.



Sent from my GT-I9506 using Tapatalk

[madjak]

I can't say I've ever noticed this with my car... I certainly haven't felt it but I'll have a look in my logs. I get the standard hunting down below 4000 rpm but thats due to cams and tune.

It might be the tyre wall moving around or some sort of oscillation in the drivetrain / crank. Did you know that awesome warbling sound that the old F1 v10's used to make is basically due to having so little MoI in the engine because they didn't run a flywheel that the engine is actually oscilating forwards and backwards between strokes. check this out:


Maybe your cars are so awesome you are getting the same effect!

[Magpie]

Had a look at 2 sets of data (Racepak and Vbox). Again I think it is more about the logging rate, however I would need to export the data and do some statistical analysis to either say there is a correlation in the bumps in the RPM to suspension movement. I picked a bumpy part of Morgan Park (layout E) as you come through the S's and it is a bit rough.

The Vbox has a higher data rate on its G force channel hence why it appears smooth and the others are spikey. The rear suspension sensors are only on the VBOX at the moment due to lack of available channels on the Racepak. I need to add a few USM's to the Racepak at some stage. I could add a TPS to the VBOX as there is a spare one in the engine bay.

VBOX
Wobble by Eipeip, on Flickr

Racepak. Also included Inj duty cycle, timing, TPS, RPM and avg EGT
Wobble_01 by Eipeip, on Flickr

[Magpie]

Another method of determining where the 'wobble' is or at least rule out something is to compare the engine RPM to the driveshaft RPM. I have no desire or inclination to do this...

[Cus]

I went a slightly different way about it, gopro on the tow hook pointed at the diff. Initial viewing indicates it moves "a bit".

Because it's Friday and I'm a party animal, I'm in the process of working out how to overlay the ECU log (If there is one, I haven't checked yet, so it's a Schrodinger's log right now) onto the video of the diff.

Then I'll upload it.

Then we party.


(Also madjak, that video is porn)

[Cus]

If you watch it in fullscreen you can see I've drawn an overlay of where some things are while the car is resting. Black lines, not terribly visible.

Sadly the log I have is for the drive i didn't video (it's hard to see if you've pressed the right button on a device that's attached to the bottom of a car!), and the video I have is the one I didn't log - however the movement seems to be less wobbly than expected.

The biggest shock is at 1:05 which is the handbrake on and me rocking back and forwards in the car a couple of times, the bottom of the sub-frame moves!
1:37 - 1:53 is the bulk of the excitement (decel, then accel doing a u-turn around the round about)

[Cus]

Had the car on a hoist today, it's *definitely* diff bushes, I can move the diff on the bushes by holding it between my thumb and one finger